Upload folder using huggingface_hub

.ipynb_checkpoints/model.safetensors.index-checkpoint.json

@@ -0,0 +1,810 @@
+{
+  "metadata": {
+    "total_size": 445909968896
+  },
+  "weight_map": {
+    "transformer.h.0.ln_attn.bias": "model-00002-of-00101.safetensors",
+    "transformer.h.0.ln_attn.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.0.ln_mlp.bias": "model-00002-of-00101.safetensors",
+    "transformer.h.0.ln_mlp.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.0.mlp.dense_4h_to_h.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.0.mlp.dense_h_to_4h.weight": "model-00001-of-00101.safetensors",
+    "transformer.h.0.self_attention.dense.weight": "model-00001-of-00101.safetensors",
+    "transformer.h.0.self_attention.query_key_value.weight": "model-00001-of-00101.safetensors",
+    "transformer.h.1.ln_attn.bias": "model-00003-of-00101.safetensors",
+    "transformer.h.1.ln_attn.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.1.ln_mlp.bias": "model-00003-of-00101.safetensors",
+    "transformer.h.1.ln_mlp.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.1.mlp.dense_4h_to_h.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.1.mlp.dense_h_to_4h.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.1.self_attention.dense.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.1.self_attention.query_key_value.weight": "model-00002-of-00101.safetensors",
+    "transformer.h.10.ln_attn.bias": "model-00012-of-00101.safetensors",
+    "transformer.h.10.ln_attn.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.10.ln_mlp.bias": "model-00012-of-00101.safetensors",
+    "transformer.h.10.ln_mlp.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.10.mlp.dense_4h_to_h.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.10.mlp.dense_h_to_4h.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.10.self_attention.dense.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.10.self_attention.query_key_value.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.11.ln_attn.bias": "model-00013-of-00101.safetensors",
+    "transformer.h.11.ln_attn.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.11.ln_mlp.bias": "model-00013-of-00101.safetensors",
+    "transformer.h.11.ln_mlp.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.11.mlp.dense_4h_to_h.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.11.mlp.dense_h_to_4h.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.11.self_attention.dense.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.11.self_attention.query_key_value.weight": "model-00012-of-00101.safetensors",
+    "transformer.h.12.ln_attn.bias": "model-00014-of-00101.safetensors",
+    "transformer.h.12.ln_attn.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.12.ln_mlp.bias": "model-00014-of-00101.safetensors",
+    "transformer.h.12.ln_mlp.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.12.mlp.dense_4h_to_h.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.12.mlp.dense_h_to_4h.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.12.self_attention.dense.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.12.self_attention.query_key_value.weight": "model-00013-of-00101.safetensors",
+    "transformer.h.13.ln_attn.bias": "model-00015-of-00101.safetensors",
+    "transformer.h.13.ln_attn.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.13.ln_mlp.bias": "model-00015-of-00101.safetensors",
+    "transformer.h.13.ln_mlp.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.13.mlp.dense_4h_to_h.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.13.mlp.dense_h_to_4h.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.13.self_attention.dense.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.13.self_attention.query_key_value.weight": "model-00014-of-00101.safetensors",
+    "transformer.h.14.ln_attn.bias": "model-00016-of-00101.safetensors",
+    "transformer.h.14.ln_attn.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.14.ln_mlp.bias": "model-00016-of-00101.safetensors",
+    "transformer.h.14.ln_mlp.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.14.mlp.dense_4h_to_h.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.14.mlp.dense_h_to_4h.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.14.self_attention.dense.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.14.self_attention.query_key_value.weight": "model-00015-of-00101.safetensors",
+    "transformer.h.15.ln_attn.bias": "model-00017-of-00101.safetensors",
+    "transformer.h.15.ln_attn.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.15.ln_mlp.bias": "model-00017-of-00101.safetensors",
+    "transformer.h.15.ln_mlp.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.15.mlp.dense_4h_to_h.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.15.mlp.dense_h_to_4h.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.15.self_attention.dense.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.15.self_attention.query_key_value.weight": "model-00016-of-00101.safetensors",
+    "transformer.h.16.ln_attn.bias": "model-00018-of-00101.safetensors",
+    "transformer.h.16.ln_attn.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.16.ln_mlp.bias": "model-00018-of-00101.safetensors",
+    "transformer.h.16.ln_mlp.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.16.mlp.dense_4h_to_h.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.16.mlp.dense_h_to_4h.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.16.self_attention.dense.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.16.self_attention.query_key_value.weight": "model-00017-of-00101.safetensors",
+    "transformer.h.17.ln_attn.bias": "model-00019-of-00101.safetensors",
+    "transformer.h.17.ln_attn.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.17.ln_mlp.bias": "model-00019-of-00101.safetensors",
+    "transformer.h.17.ln_mlp.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.17.mlp.dense_4h_to_h.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.17.mlp.dense_h_to_4h.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.17.self_attention.dense.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.17.self_attention.query_key_value.weight": "model-00018-of-00101.safetensors",
+    "transformer.h.18.ln_attn.bias": "model-00020-of-00101.safetensors",
+    "transformer.h.18.ln_attn.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.18.ln_mlp.bias": "model-00020-of-00101.safetensors",
+    "transformer.h.18.ln_mlp.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.18.mlp.dense_4h_to_h.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.18.mlp.dense_h_to_4h.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.18.self_attention.dense.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.18.self_attention.query_key_value.weight": "model-00019-of-00101.safetensors",
+    "transformer.h.19.ln_attn.bias": "model-00021-of-00101.safetensors",
+    "transformer.h.19.ln_attn.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.19.ln_mlp.bias": "model-00021-of-00101.safetensors",
+    "transformer.h.19.ln_mlp.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.19.mlp.dense_4h_to_h.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.19.mlp.dense_h_to_4h.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.19.self_attention.dense.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.19.self_attention.query_key_value.weight": "model-00020-of-00101.safetensors",
+    "transformer.h.2.ln_attn.bias": "model-00004-of-00101.safetensors",
+    "transformer.h.2.ln_attn.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.2.ln_mlp.bias": "model-00004-of-00101.safetensors",
+    "transformer.h.2.ln_mlp.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.2.mlp.dense_4h_to_h.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.2.mlp.dense_h_to_4h.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.2.self_attention.dense.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.2.self_attention.query_key_value.weight": "model-00003-of-00101.safetensors",
+    "transformer.h.20.ln_attn.bias": "model-00022-of-00101.safetensors",
+    "transformer.h.20.ln_attn.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.20.ln_mlp.bias": "model-00022-of-00101.safetensors",
+    "transformer.h.20.ln_mlp.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.20.mlp.dense_4h_to_h.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.20.mlp.dense_h_to_4h.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.20.self_attention.dense.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.20.self_attention.query_key_value.weight": "model-00021-of-00101.safetensors",
+    "transformer.h.21.ln_attn.bias": "model-00023-of-00101.safetensors",
+    "transformer.h.21.ln_attn.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.21.ln_mlp.bias": "model-00023-of-00101.safetensors",
+    "transformer.h.21.ln_mlp.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.21.mlp.dense_4h_to_h.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.21.mlp.dense_h_to_4h.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.21.self_attention.dense.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.21.self_attention.query_key_value.weight": "model-00022-of-00101.safetensors",
+    "transformer.h.22.ln_attn.bias": "model-00024-of-00101.safetensors",
+    "transformer.h.22.ln_attn.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.22.ln_mlp.bias": "model-00024-of-00101.safetensors",
+    "transformer.h.22.ln_mlp.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.22.mlp.dense_4h_to_h.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.22.mlp.dense_h_to_4h.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.22.self_attention.dense.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.22.self_attention.query_key_value.weight": "model-00023-of-00101.safetensors",
+    "transformer.h.23.ln_attn.bias": "model-00025-of-00101.safetensors",
+    "transformer.h.23.ln_attn.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.23.ln_mlp.bias": "model-00025-of-00101.safetensors",
+    "transformer.h.23.ln_mlp.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.23.mlp.dense_4h_to_h.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.23.mlp.dense_h_to_4h.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.23.self_attention.dense.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.23.self_attention.query_key_value.weight": "model-00024-of-00101.safetensors",
+    "transformer.h.24.ln_attn.bias": "model-00026-of-00101.safetensors",
+    "transformer.h.24.ln_attn.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.24.ln_mlp.bias": "model-00026-of-00101.safetensors",
+    "transformer.h.24.ln_mlp.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.24.mlp.dense_4h_to_h.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.24.mlp.dense_h_to_4h.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.24.self_attention.dense.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.24.self_attention.query_key_value.weight": "model-00025-of-00101.safetensors",
+    "transformer.h.25.ln_attn.bias": "model-00027-of-00101.safetensors",
+    "transformer.h.25.ln_attn.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.25.ln_mlp.bias": "model-00027-of-00101.safetensors",
+    "transformer.h.25.ln_mlp.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.25.mlp.dense_4h_to_h.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.25.mlp.dense_h_to_4h.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.25.self_attention.dense.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.25.self_attention.query_key_value.weight": "model-00026-of-00101.safetensors",
+    "transformer.h.26.ln_attn.bias": "model-00028-of-00101.safetensors",
+    "transformer.h.26.ln_attn.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.26.ln_mlp.bias": "model-00028-of-00101.safetensors",
+    "transformer.h.26.ln_mlp.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.26.mlp.dense_4h_to_h.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.26.mlp.dense_h_to_4h.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.26.self_attention.dense.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.26.self_attention.query_key_value.weight": "model-00027-of-00101.safetensors",
+    "transformer.h.27.ln_attn.bias": "model-00029-of-00101.safetensors",
+    "transformer.h.27.ln_attn.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.27.ln_mlp.bias": "model-00029-of-00101.safetensors",
+    "transformer.h.27.ln_mlp.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.27.mlp.dense_4h_to_h.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.27.mlp.dense_h_to_4h.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.27.self_attention.dense.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.27.self_attention.query_key_value.weight": "model-00028-of-00101.safetensors",
+    "transformer.h.28.ln_attn.bias": "model-00030-of-00101.safetensors",
+    "transformer.h.28.ln_attn.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.28.ln_mlp.bias": "model-00030-of-00101.safetensors",
+    "transformer.h.28.ln_mlp.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.28.mlp.dense_4h_to_h.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.28.mlp.dense_h_to_4h.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.28.self_attention.dense.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.28.self_attention.query_key_value.weight": "model-00029-of-00101.safetensors",
+    "transformer.h.29.ln_attn.bias": "model-00031-of-00101.safetensors",
+    "transformer.h.29.ln_attn.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.29.ln_mlp.bias": "model-00031-of-00101.safetensors",
+    "transformer.h.29.ln_mlp.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.29.mlp.dense_4h_to_h.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.29.mlp.dense_h_to_4h.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.29.self_attention.dense.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.29.self_attention.query_key_value.weight": "model-00030-of-00101.safetensors",
+    "transformer.h.3.ln_attn.bias": "model-00005-of-00101.safetensors",
+    "transformer.h.3.ln_attn.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.3.ln_mlp.bias": "model-00005-of-00101.safetensors",
+    "transformer.h.3.ln_mlp.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.3.mlp.dense_4h_to_h.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.3.mlp.dense_h_to_4h.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.3.self_attention.dense.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.3.self_attention.query_key_value.weight": "model-00004-of-00101.safetensors",
+    "transformer.h.30.ln_attn.bias": "model-00032-of-00101.safetensors",
+    "transformer.h.30.ln_attn.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.30.ln_mlp.bias": "model-00032-of-00101.safetensors",
+    "transformer.h.30.ln_mlp.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.30.mlp.dense_4h_to_h.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.30.mlp.dense_h_to_4h.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.30.self_attention.dense.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.30.self_attention.query_key_value.weight": "model-00031-of-00101.safetensors",
+    "transformer.h.31.ln_attn.bias": "model-00033-of-00101.safetensors",
+    "transformer.h.31.ln_attn.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.31.ln_mlp.bias": "model-00033-of-00101.safetensors",
+    "transformer.h.31.ln_mlp.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.31.mlp.dense_4h_to_h.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.31.mlp.dense_h_to_4h.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.31.self_attention.dense.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.31.self_attention.query_key_value.weight": "model-00032-of-00101.safetensors",
+    "transformer.h.32.ln_attn.bias": "model-00034-of-00101.safetensors",
+    "transformer.h.32.ln_attn.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.32.ln_mlp.bias": "model-00034-of-00101.safetensors",
+    "transformer.h.32.ln_mlp.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.32.mlp.dense_4h_to_h.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.32.mlp.dense_h_to_4h.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.32.self_attention.dense.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.32.self_attention.query_key_value.weight": "model-00033-of-00101.safetensors",
+    "transformer.h.33.ln_attn.bias": "model-00035-of-00101.safetensors",
+    "transformer.h.33.ln_attn.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.33.ln_mlp.bias": "model-00035-of-00101.safetensors",
+    "transformer.h.33.ln_mlp.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.33.mlp.dense_4h_to_h.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.33.mlp.dense_h_to_4h.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.33.self_attention.dense.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.33.self_attention.query_key_value.weight": "model-00034-of-00101.safetensors",
+    "transformer.h.34.ln_attn.bias": "model-00036-of-00101.safetensors",
+    "transformer.h.34.ln_attn.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.34.ln_mlp.bias": "model-00036-of-00101.safetensors",
+    "transformer.h.34.ln_mlp.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.34.mlp.dense_4h_to_h.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.34.mlp.dense_h_to_4h.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.34.self_attention.dense.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.34.self_attention.query_key_value.weight": "model-00035-of-00101.safetensors",
+    "transformer.h.35.ln_attn.bias": "model-00037-of-00101.safetensors",
+    "transformer.h.35.ln_attn.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.35.ln_mlp.bias": "model-00037-of-00101.safetensors",
+    "transformer.h.35.ln_mlp.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.35.mlp.dense_4h_to_h.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.35.mlp.dense_h_to_4h.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.35.self_attention.dense.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.35.self_attention.query_key_value.weight": "model-00036-of-00101.safetensors",
+    "transformer.h.36.ln_attn.bias": "model-00038-of-00101.safetensors",
+    "transformer.h.36.ln_attn.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.36.ln_mlp.bias": "model-00038-of-00101.safetensors",
+    "transformer.h.36.ln_mlp.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.36.mlp.dense_4h_to_h.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.36.mlp.dense_h_to_4h.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.36.self_attention.dense.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.36.self_attention.query_key_value.weight": "model-00037-of-00101.safetensors",
+    "transformer.h.37.ln_attn.bias": "model-00039-of-00101.safetensors",
+    "transformer.h.37.ln_attn.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.37.ln_mlp.bias": "model-00039-of-00101.safetensors",
+    "transformer.h.37.ln_mlp.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.37.mlp.dense_4h_to_h.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.37.mlp.dense_h_to_4h.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.37.self_attention.dense.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.37.self_attention.query_key_value.weight": "model-00038-of-00101.safetensors",
+    "transformer.h.38.ln_attn.bias": "model-00040-of-00101.safetensors",
+    "transformer.h.38.ln_attn.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.38.ln_mlp.bias": "model-00040-of-00101.safetensors",
+    "transformer.h.38.ln_mlp.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.38.mlp.dense_4h_to_h.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.38.mlp.dense_h_to_4h.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.38.self_attention.dense.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.38.self_attention.query_key_value.weight": "model-00039-of-00101.safetensors",
+    "transformer.h.39.ln_attn.bias": "model-00041-of-00101.safetensors",
+    "transformer.h.39.ln_attn.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.39.ln_mlp.bias": "model-00041-of-00101.safetensors",
+    "transformer.h.39.ln_mlp.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.39.mlp.dense_4h_to_h.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.39.mlp.dense_h_to_4h.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.39.self_attention.dense.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.39.self_attention.query_key_value.weight": "model-00040-of-00101.safetensors",
+    "transformer.h.4.ln_attn.bias": "model-00006-of-00101.safetensors",
+    "transformer.h.4.ln_attn.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.4.ln_mlp.bias": "model-00006-of-00101.safetensors",
+    "transformer.h.4.ln_mlp.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.4.mlp.dense_4h_to_h.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.4.mlp.dense_h_to_4h.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.4.self_attention.dense.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.4.self_attention.query_key_value.weight": "model-00005-of-00101.safetensors",
+    "transformer.h.40.ln_attn.bias": "model-00042-of-00101.safetensors",
+    "transformer.h.40.ln_attn.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.40.ln_mlp.bias": "model-00042-of-00101.safetensors",
+    "transformer.h.40.ln_mlp.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.40.mlp.dense_4h_to_h.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.40.mlp.dense_h_to_4h.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.40.self_attention.dense.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.40.self_attention.query_key_value.weight": "model-00041-of-00101.safetensors",
+    "transformer.h.41.ln_attn.bias": "model-00043-of-00101.safetensors",
+    "transformer.h.41.ln_attn.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.41.ln_mlp.bias": "model-00043-of-00101.safetensors",
+    "transformer.h.41.ln_mlp.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.41.mlp.dense_4h_to_h.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.41.mlp.dense_h_to_4h.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.41.self_attention.dense.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.41.self_attention.query_key_value.weight": "model-00042-of-00101.safetensors",
+    "transformer.h.42.ln_attn.bias": "model-00044-of-00101.safetensors",
+    "transformer.h.42.ln_attn.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.42.ln_mlp.bias": "model-00044-of-00101.safetensors",
+    "transformer.h.42.ln_mlp.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.42.mlp.dense_4h_to_h.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.42.mlp.dense_h_to_4h.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.42.self_attention.dense.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.42.self_attention.query_key_value.weight": "model-00043-of-00101.safetensors",
+    "transformer.h.43.ln_attn.bias": "model-00045-of-00101.safetensors",
+    "transformer.h.43.ln_attn.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.43.ln_mlp.bias": "model-00045-of-00101.safetensors",
+    "transformer.h.43.ln_mlp.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.43.mlp.dense_4h_to_h.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.43.mlp.dense_h_to_4h.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.43.self_attention.dense.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.43.self_attention.query_key_value.weight": "model-00044-of-00101.safetensors",
+    "transformer.h.44.ln_attn.bias": "model-00046-of-00101.safetensors",
+    "transformer.h.44.ln_attn.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.44.ln_mlp.bias": "model-00046-of-00101.safetensors",
+    "transformer.h.44.ln_mlp.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.44.mlp.dense_4h_to_h.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.44.mlp.dense_h_to_4h.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.44.self_attention.dense.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.44.self_attention.query_key_value.weight": "model-00045-of-00101.safetensors",
+    "transformer.h.45.ln_attn.bias": "model-00047-of-00101.safetensors",
+    "transformer.h.45.ln_attn.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.45.ln_mlp.bias": "model-00047-of-00101.safetensors",
+    "transformer.h.45.ln_mlp.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.45.mlp.dense_4h_to_h.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.45.mlp.dense_h_to_4h.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.45.self_attention.dense.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.45.self_attention.query_key_value.weight": "model-00046-of-00101.safetensors",
+    "transformer.h.46.ln_attn.bias": "model-00048-of-00101.safetensors",
+    "transformer.h.46.ln_attn.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.46.ln_mlp.bias": "model-00048-of-00101.safetensors",
+    "transformer.h.46.ln_mlp.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.46.mlp.dense_4h_to_h.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.46.mlp.dense_h_to_4h.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.46.self_attention.dense.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.46.self_attention.query_key_value.weight": "model-00047-of-00101.safetensors",
+    "transformer.h.47.ln_attn.bias": "model-00049-of-00101.safetensors",
+    "transformer.h.47.ln_attn.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.47.ln_mlp.bias": "model-00049-of-00101.safetensors",
+    "transformer.h.47.ln_mlp.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.47.mlp.dense_4h_to_h.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.47.mlp.dense_h_to_4h.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.47.self_attention.dense.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.47.self_attention.query_key_value.weight": "model-00048-of-00101.safetensors",
+    "transformer.h.48.ln_attn.bias": "model-00050-of-00101.safetensors",
+    "transformer.h.48.ln_attn.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.48.ln_mlp.bias": "model-00050-of-00101.safetensors",
+    "transformer.h.48.ln_mlp.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.48.mlp.dense_4h_to_h.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.48.mlp.dense_h_to_4h.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.48.self_attention.dense.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.48.self_attention.query_key_value.weight": "model-00049-of-00101.safetensors",
+    "transformer.h.49.ln_attn.bias": "model-00051-of-00101.safetensors",
+    "transformer.h.49.ln_attn.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.49.ln_mlp.bias": "model-00051-of-00101.safetensors",
+    "transformer.h.49.ln_mlp.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.49.mlp.dense_4h_to_h.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.49.mlp.dense_h_to_4h.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.49.self_attention.dense.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.49.self_attention.query_key_value.weight": "model-00050-of-00101.safetensors",
+    "transformer.h.5.ln_attn.bias": "model-00007-of-00101.safetensors",
+    "transformer.h.5.ln_attn.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.5.ln_mlp.bias": "model-00007-of-00101.safetensors",
+    "transformer.h.5.ln_mlp.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.5.mlp.dense_4h_to_h.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.5.mlp.dense_h_to_4h.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.5.self_attention.dense.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.5.self_attention.query_key_value.weight": "model-00006-of-00101.safetensors",
+    "transformer.h.50.ln_attn.bias": "model-00052-of-00101.safetensors",
+    "transformer.h.50.ln_attn.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.50.ln_mlp.bias": "model-00052-of-00101.safetensors",
+    "transformer.h.50.ln_mlp.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.50.mlp.dense_4h_to_h.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.50.mlp.dense_h_to_4h.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.50.self_attention.dense.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.50.self_attention.query_key_value.weight": "model-00051-of-00101.safetensors",
+    "transformer.h.51.ln_attn.bias": "model-00053-of-00101.safetensors",
+    "transformer.h.51.ln_attn.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.51.ln_mlp.bias": "model-00053-of-00101.safetensors",
+    "transformer.h.51.ln_mlp.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.51.mlp.dense_4h_to_h.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.51.mlp.dense_h_to_4h.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.51.self_attention.dense.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.51.self_attention.query_key_value.weight": "model-00052-of-00101.safetensors",
+    "transformer.h.52.ln_attn.bias": "model-00054-of-00101.safetensors",
+    "transformer.h.52.ln_attn.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.52.ln_mlp.bias": "model-00054-of-00101.safetensors",
+    "transformer.h.52.ln_mlp.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.52.mlp.dense_4h_to_h.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.52.mlp.dense_h_to_4h.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.52.self_attention.dense.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.52.self_attention.query_key_value.weight": "model-00053-of-00101.safetensors",
+    "transformer.h.53.ln_attn.bias": "model-00055-of-00101.safetensors",
+    "transformer.h.53.ln_attn.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.53.ln_mlp.bias": "model-00055-of-00101.safetensors",
+    "transformer.h.53.ln_mlp.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.53.mlp.dense_4h_to_h.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.53.mlp.dense_h_to_4h.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.53.self_attention.dense.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.53.self_attention.query_key_value.weight": "model-00054-of-00101.safetensors",
+    "transformer.h.54.ln_attn.bias": "model-00056-of-00101.safetensors",
+    "transformer.h.54.ln_attn.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.54.ln_mlp.bias": "model-00056-of-00101.safetensors",
+    "transformer.h.54.ln_mlp.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.54.mlp.dense_4h_to_h.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.54.mlp.dense_h_to_4h.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.54.self_attention.dense.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.54.self_attention.query_key_value.weight": "model-00055-of-00101.safetensors",
+    "transformer.h.55.ln_attn.bias": "model-00057-of-00101.safetensors",
+    "transformer.h.55.ln_attn.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.55.ln_mlp.bias": "model-00057-of-00101.safetensors",
+    "transformer.h.55.ln_mlp.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.55.mlp.dense_4h_to_h.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.55.mlp.dense_h_to_4h.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.55.self_attention.dense.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.55.self_attention.query_key_value.weight": "model-00056-of-00101.safetensors",
+    "transformer.h.56.ln_attn.bias": "model-00058-of-00101.safetensors",
+    "transformer.h.56.ln_attn.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.56.ln_mlp.bias": "model-00058-of-00101.safetensors",
+    "transformer.h.56.ln_mlp.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.56.mlp.dense_4h_to_h.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.56.mlp.dense_h_to_4h.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.56.self_attention.dense.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.56.self_attention.query_key_value.weight": "model-00057-of-00101.safetensors",
+    "transformer.h.57.ln_attn.bias": "model-00059-of-00101.safetensors",
+    "transformer.h.57.ln_attn.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.57.ln_mlp.bias": "model-00059-of-00101.safetensors",
+    "transformer.h.57.ln_mlp.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.57.mlp.dense_4h_to_h.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.57.mlp.dense_h_to_4h.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.57.self_attention.dense.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.57.self_attention.query_key_value.weight": "model-00058-of-00101.safetensors",
+    "transformer.h.58.ln_attn.bias": "model-00060-of-00101.safetensors",
+    "transformer.h.58.ln_attn.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.58.ln_mlp.bias": "model-00060-of-00101.safetensors",
+    "transformer.h.58.ln_mlp.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.58.mlp.dense_4h_to_h.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.58.mlp.dense_h_to_4h.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.58.self_attention.dense.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.58.self_attention.query_key_value.weight": "model-00059-of-00101.safetensors",
+    "transformer.h.59.ln_attn.bias": "model-00061-of-00101.safetensors",
+    "transformer.h.59.ln_attn.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.59.ln_mlp.bias": "model-00061-of-00101.safetensors",
+    "transformer.h.59.ln_mlp.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.59.mlp.dense_4h_to_h.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.59.mlp.dense_h_to_4h.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.59.self_attention.dense.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.59.self_attention.query_key_value.weight": "model-00060-of-00101.safetensors",
+    "transformer.h.6.ln_attn.bias": "model-00008-of-00101.safetensors",
+    "transformer.h.6.ln_attn.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.6.ln_mlp.bias": "model-00008-of-00101.safetensors",
+    "transformer.h.6.ln_mlp.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.6.mlp.dense_4h_to_h.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.6.mlp.dense_h_to_4h.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.6.self_attention.dense.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.6.self_attention.query_key_value.weight": "model-00007-of-00101.safetensors",
+    "transformer.h.60.ln_attn.bias": "model-00062-of-00101.safetensors",
+    "transformer.h.60.ln_attn.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.60.ln_mlp.bias": "model-00062-of-00101.safetensors",
+    "transformer.h.60.ln_mlp.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.60.mlp.dense_4h_to_h.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.60.mlp.dense_h_to_4h.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.60.self_attention.dense.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.60.self_attention.query_key_value.weight": "model-00061-of-00101.safetensors",
+    "transformer.h.61.ln_attn.bias": "model-00063-of-00101.safetensors",
+    "transformer.h.61.ln_attn.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.61.ln_mlp.bias": "model-00063-of-00101.safetensors",
+    "transformer.h.61.ln_mlp.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.61.mlp.dense_4h_to_h.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.61.mlp.dense_h_to_4h.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.61.self_attention.dense.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.61.self_attention.query_key_value.weight": "model-00062-of-00101.safetensors",
+    "transformer.h.62.ln_attn.bias": "model-00064-of-00101.safetensors",
+    "transformer.h.62.ln_attn.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.62.ln_mlp.bias": "model-00064-of-00101.safetensors",
+    "transformer.h.62.ln_mlp.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.62.mlp.dense_4h_to_h.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.62.mlp.dense_h_to_4h.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.62.self_attention.dense.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.62.self_attention.query_key_value.weight": "model-00063-of-00101.safetensors",
+    "transformer.h.63.ln_attn.bias": "model-00065-of-00101.safetensors",
+    "transformer.h.63.ln_attn.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.63.ln_mlp.bias": "model-00065-of-00101.safetensors",
+    "transformer.h.63.ln_mlp.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.63.mlp.dense_4h_to_h.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.63.mlp.dense_h_to_4h.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.63.self_attention.dense.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.63.self_attention.query_key_value.weight": "model-00064-of-00101.safetensors",
+    "transformer.h.64.ln_attn.bias": "model-00066-of-00101.safetensors",
+    "transformer.h.64.ln_attn.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.64.ln_mlp.bias": "model-00066-of-00101.safetensors",
+    "transformer.h.64.ln_mlp.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.64.mlp.dense_4h_to_h.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.64.mlp.dense_h_to_4h.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.64.self_attention.dense.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.64.self_attention.query_key_value.weight": "model-00065-of-00101.safetensors",
+    "transformer.h.65.ln_attn.bias": "model-00067-of-00101.safetensors",
+    "transformer.h.65.ln_attn.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.65.ln_mlp.bias": "model-00067-of-00101.safetensors",
+    "transformer.h.65.ln_mlp.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.65.mlp.dense_4h_to_h.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.65.mlp.dense_h_to_4h.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.65.self_attention.dense.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.65.self_attention.query_key_value.weight": "model-00066-of-00101.safetensors",
+    "transformer.h.66.ln_attn.bias": "model-00068-of-00101.safetensors",
+    "transformer.h.66.ln_attn.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.66.ln_mlp.bias": "model-00068-of-00101.safetensors",
+    "transformer.h.66.ln_mlp.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.66.mlp.dense_4h_to_h.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.66.mlp.dense_h_to_4h.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.66.self_attention.dense.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.66.self_attention.query_key_value.weight": "model-00067-of-00101.safetensors",
+    "transformer.h.67.ln_attn.bias": "model-00069-of-00101.safetensors",
+    "transformer.h.67.ln_attn.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.67.ln_mlp.bias": "model-00069-of-00101.safetensors",
+    "transformer.h.67.ln_mlp.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.67.mlp.dense_4h_to_h.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.67.mlp.dense_h_to_4h.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.67.self_attention.dense.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.67.self_attention.query_key_value.weight": "model-00068-of-00101.safetensors",
+    "transformer.h.68.ln_attn.bias": "model-00070-of-00101.safetensors",
+    "transformer.h.68.ln_attn.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.68.ln_mlp.bias": "model-00070-of-00101.safetensors",
+    "transformer.h.68.ln_mlp.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.68.mlp.dense_4h_to_h.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.68.mlp.dense_h_to_4h.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.68.self_attention.dense.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.68.self_attention.query_key_value.weight": "model-00069-of-00101.safetensors",
+    "transformer.h.69.ln_attn.bias": "model-00071-of-00101.safetensors",
+    "transformer.h.69.ln_attn.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.69.ln_mlp.bias": "model-00071-of-00101.safetensors",
+    "transformer.h.69.ln_mlp.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.69.mlp.dense_4h_to_h.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.69.mlp.dense_h_to_4h.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.69.self_attention.dense.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.69.self_attention.query_key_value.weight": "model-00070-of-00101.safetensors",
+    "transformer.h.7.ln_attn.bias": "model-00009-of-00101.safetensors",
+    "transformer.h.7.ln_attn.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.7.ln_mlp.bias": "model-00009-of-00101.safetensors",
+    "transformer.h.7.ln_mlp.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.7.mlp.dense_4h_to_h.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.7.mlp.dense_h_to_4h.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.7.self_attention.dense.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.7.self_attention.query_key_value.weight": "model-00008-of-00101.safetensors",
+    "transformer.h.70.ln_attn.bias": "model-00072-of-00101.safetensors",
+    "transformer.h.70.ln_attn.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.70.ln_mlp.bias": "model-00072-of-00101.safetensors",
+    "transformer.h.70.ln_mlp.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.70.mlp.dense_4h_to_h.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.70.mlp.dense_h_to_4h.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.70.self_attention.dense.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.70.self_attention.query_key_value.weight": "model-00071-of-00101.safetensors",
+    "transformer.h.71.ln_attn.bias": "model-00073-of-00101.safetensors",
+    "transformer.h.71.ln_attn.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.71.ln_mlp.bias": "model-00073-of-00101.safetensors",
+    "transformer.h.71.ln_mlp.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.71.mlp.dense_4h_to_h.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.71.mlp.dense_h_to_4h.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.71.self_attention.dense.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.71.self_attention.query_key_value.weight": "model-00072-of-00101.safetensors",
+    "transformer.h.72.ln_attn.bias": "model-00074-of-00101.safetensors",
+    "transformer.h.72.ln_attn.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.72.ln_mlp.bias": "model-00074-of-00101.safetensors",
+    "transformer.h.72.ln_mlp.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.72.mlp.dense_4h_to_h.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.72.mlp.dense_h_to_4h.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.72.self_attention.dense.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.72.self_attention.query_key_value.weight": "model-00073-of-00101.safetensors",
+    "transformer.h.73.ln_attn.bias": "model-00075-of-00101.safetensors",
+    "transformer.h.73.ln_attn.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.73.ln_mlp.bias": "model-00075-of-00101.safetensors",
+    "transformer.h.73.ln_mlp.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.73.mlp.dense_4h_to_h.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.73.mlp.dense_h_to_4h.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.73.self_attention.dense.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.73.self_attention.query_key_value.weight": "model-00074-of-00101.safetensors",
+    "transformer.h.74.ln_attn.bias": "model-00076-of-00101.safetensors",
+    "transformer.h.74.ln_attn.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.74.ln_mlp.bias": "model-00076-of-00101.safetensors",
+    "transformer.h.74.ln_mlp.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.74.mlp.dense_4h_to_h.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.74.mlp.dense_h_to_4h.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.74.self_attention.dense.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.74.self_attention.query_key_value.weight": "model-00075-of-00101.safetensors",
+    "transformer.h.75.ln_attn.bias": "model-00077-of-00101.safetensors",
+    "transformer.h.75.ln_attn.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.75.ln_mlp.bias": "model-00077-of-00101.safetensors",
+    "transformer.h.75.ln_mlp.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.75.mlp.dense_4h_to_h.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.75.mlp.dense_h_to_4h.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.75.self_attention.dense.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.75.self_attention.query_key_value.weight": "model-00076-of-00101.safetensors",
+    "transformer.h.76.ln_attn.bias": "model-00078-of-00101.safetensors",
+    "transformer.h.76.ln_attn.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.76.ln_mlp.bias": "model-00078-of-00101.safetensors",
+    "transformer.h.76.ln_mlp.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.76.mlp.dense_4h_to_h.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.76.mlp.dense_h_to_4h.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.76.self_attention.dense.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.76.self_attention.query_key_value.weight": "model-00077-of-00101.safetensors",
+    "transformer.h.77.ln_attn.bias": "model-00079-of-00101.safetensors",
+    "transformer.h.77.ln_attn.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.77.ln_mlp.bias": "model-00079-of-00101.safetensors",
+    "transformer.h.77.ln_mlp.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.77.mlp.dense_4h_to_h.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.77.mlp.dense_h_to_4h.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.77.self_attention.dense.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.77.self_attention.query_key_value.weight": "model-00078-of-00101.safetensors",
+    "transformer.h.78.ln_attn.bias": "model-00080-of-00101.safetensors",
+    "transformer.h.78.ln_attn.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.78.ln_mlp.bias": "model-00080-of-00101.safetensors",
+    "transformer.h.78.ln_mlp.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.78.mlp.dense_4h_to_h.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.78.mlp.dense_h_to_4h.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.78.self_attention.dense.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.78.self_attention.query_key_value.weight": "model-00079-of-00101.safetensors",
+    "transformer.h.79.ln_attn.bias": "model-00081-of-00101.safetensors",
+    "transformer.h.79.ln_attn.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.79.ln_mlp.bias": "model-00081-of-00101.safetensors",
+    "transformer.h.79.ln_mlp.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.79.mlp.dense_4h_to_h.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.79.mlp.dense_h_to_4h.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.79.self_attention.dense.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.79.self_attention.query_key_value.weight": "model-00080-of-00101.safetensors",
+    "transformer.h.8.ln_attn.bias": "model-00010-of-00101.safetensors",
+    "transformer.h.8.ln_attn.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.8.ln_mlp.bias": "model-00010-of-00101.safetensors",
+    "transformer.h.8.ln_mlp.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.8.mlp.dense_4h_to_h.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.8.mlp.dense_h_to_4h.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.8.self_attention.dense.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.8.self_attention.query_key_value.weight": "model-00009-of-00101.safetensors",
+    "transformer.h.80.ln_attn.bias": "model-00082-of-00101.safetensors",
+    "transformer.h.80.ln_attn.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.80.ln_mlp.bias": "model-00082-of-00101.safetensors",
+    "transformer.h.80.ln_mlp.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.80.mlp.dense_4h_to_h.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.80.mlp.dense_h_to_4h.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.80.self_attention.dense.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.80.self_attention.query_key_value.weight": "model-00081-of-00101.safetensors",
+    "transformer.h.81.ln_attn.bias": "model-00083-of-00101.safetensors",
+    "transformer.h.81.ln_attn.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.81.ln_mlp.bias": "model-00083-of-00101.safetensors",
+    "transformer.h.81.ln_mlp.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.81.mlp.dense_4h_to_h.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.81.mlp.dense_h_to_4h.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.81.self_attention.dense.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.81.self_attention.query_key_value.weight": "model-00082-of-00101.safetensors",
+    "transformer.h.82.ln_attn.bias": "model-00084-of-00101.safetensors",
+    "transformer.h.82.ln_attn.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.82.ln_mlp.bias": "model-00084-of-00101.safetensors",
+    "transformer.h.82.ln_mlp.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.82.mlp.dense_4h_to_h.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.82.mlp.dense_h_to_4h.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.82.self_attention.dense.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.82.self_attention.query_key_value.weight": "model-00083-of-00101.safetensors",
+    "transformer.h.83.ln_attn.bias": "model-00085-of-00101.safetensors",
+    "transformer.h.83.ln_attn.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.83.ln_mlp.bias": "model-00085-of-00101.safetensors",
+    "transformer.h.83.ln_mlp.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.83.mlp.dense_4h_to_h.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.83.mlp.dense_h_to_4h.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.83.self_attention.dense.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.83.self_attention.query_key_value.weight": "model-00084-of-00101.safetensors",
+    "transformer.h.84.ln_attn.bias": "model-00086-of-00101.safetensors",
+    "transformer.h.84.ln_attn.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.84.ln_mlp.bias": "model-00086-of-00101.safetensors",
+    "transformer.h.84.ln_mlp.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.84.mlp.dense_4h_to_h.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.84.mlp.dense_h_to_4h.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.84.self_attention.dense.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.84.self_attention.query_key_value.weight": "model-00085-of-00101.safetensors",
+    "transformer.h.85.ln_attn.bias": "model-00087-of-00101.safetensors",
+    "transformer.h.85.ln_attn.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.85.ln_mlp.bias": "model-00087-of-00101.safetensors",
+    "transformer.h.85.ln_mlp.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.85.mlp.dense_4h_to_h.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.85.mlp.dense_h_to_4h.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.85.self_attention.dense.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.85.self_attention.query_key_value.weight": "model-00086-of-00101.safetensors",
+    "transformer.h.86.ln_attn.bias": "model-00088-of-00101.safetensors",
+    "transformer.h.86.ln_attn.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.86.ln_mlp.bias": "model-00088-of-00101.safetensors",
+    "transformer.h.86.ln_mlp.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.86.mlp.dense_4h_to_h.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.86.mlp.dense_h_to_4h.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.86.self_attention.dense.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.86.self_attention.query_key_value.weight": "model-00087-of-00101.safetensors",
+    "transformer.h.87.ln_attn.bias": "model-00089-of-00101.safetensors",
+    "transformer.h.87.ln_attn.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.87.ln_mlp.bias": "model-00089-of-00101.safetensors",
+    "transformer.h.87.ln_mlp.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.87.mlp.dense_4h_to_h.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.87.mlp.dense_h_to_4h.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.87.self_attention.dense.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.87.self_attention.query_key_value.weight": "model-00088-of-00101.safetensors",
+    "transformer.h.88.ln_attn.bias": "model-00090-of-00101.safetensors",
+    "transformer.h.88.ln_attn.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.88.ln_mlp.bias": "model-00090-of-00101.safetensors",
+    "transformer.h.88.ln_mlp.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.88.mlp.dense_4h_to_h.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.88.mlp.dense_h_to_4h.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.88.self_attention.dense.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.88.self_attention.query_key_value.weight": "model-00089-of-00101.safetensors",
+    "transformer.h.89.ln_attn.bias": "model-00091-of-00101.safetensors",
+    "transformer.h.89.ln_attn.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.89.ln_mlp.bias": "model-00091-of-00101.safetensors",
+    "transformer.h.89.ln_mlp.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.89.mlp.dense_4h_to_h.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.89.mlp.dense_h_to_4h.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.89.self_attention.dense.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.89.self_attention.query_key_value.weight": "model-00090-of-00101.safetensors",
+    "transformer.h.9.ln_attn.bias": "model-00011-of-00101.safetensors",
+    "transformer.h.9.ln_attn.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.9.ln_mlp.bias": "model-00011-of-00101.safetensors",
+    "transformer.h.9.ln_mlp.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.9.mlp.dense_4h_to_h.weight": "model-00011-of-00101.safetensors",
+    "transformer.h.9.mlp.dense_h_to_4h.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.9.self_attention.dense.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.9.self_attention.query_key_value.weight": "model-00010-of-00101.safetensors",
+    "transformer.h.90.ln_attn.bias": "model-00092-of-00101.safetensors",
+    "transformer.h.90.ln_attn.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.90.ln_mlp.bias": "model-00092-of-00101.safetensors",
+    "transformer.h.90.ln_mlp.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.90.mlp.dense_4h_to_h.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.90.mlp.dense_h_to_4h.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.90.self_attention.dense.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.90.self_attention.query_key_value.weight": "model-00091-of-00101.safetensors",
+    "transformer.h.91.ln_attn.bias": "model-00093-of-00101.safetensors",
+    "transformer.h.91.ln_attn.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.91.ln_mlp.bias": "model-00093-of-00101.safetensors",
+    "transformer.h.91.ln_mlp.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.91.mlp.dense_4h_to_h.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.91.mlp.dense_h_to_4h.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.91.self_attention.dense.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.91.self_attention.query_key_value.weight": "model-00092-of-00101.safetensors",
+    "transformer.h.92.ln_attn.bias": "model-00094-of-00101.safetensors",
+    "transformer.h.92.ln_attn.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.92.ln_mlp.bias": "model-00094-of-00101.safetensors",
+    "transformer.h.92.ln_mlp.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.92.mlp.dense_4h_to_h.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.92.mlp.dense_h_to_4h.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.92.self_attention.dense.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.92.self_attention.query_key_value.weight": "model-00093-of-00101.safetensors",
+    "transformer.h.93.ln_attn.bias": "model-00095-of-00101.safetensors",
+    "transformer.h.93.ln_attn.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.93.ln_mlp.bias": "model-00095-of-00101.safetensors",
+    "transformer.h.93.ln_mlp.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.93.mlp.dense_4h_to_h.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.93.mlp.dense_h_to_4h.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.93.self_attention.dense.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.93.self_attention.query_key_value.weight": "model-00094-of-00101.safetensors",
+    "transformer.h.94.ln_attn.bias": "model-00096-of-00101.safetensors",
+    "transformer.h.94.ln_attn.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.94.ln_mlp.bias": "model-00096-of-00101.safetensors",
+    "transformer.h.94.ln_mlp.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.94.mlp.dense_4h_to_h.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.94.mlp.dense_h_to_4h.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.94.self_attention.dense.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.94.self_attention.query_key_value.weight": "model-00095-of-00101.safetensors",
+    "transformer.h.95.ln_attn.bias": "model-00097-of-00101.safetensors",
+    "transformer.h.95.ln_attn.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.95.ln_mlp.bias": "model-00097-of-00101.safetensors",
+    "transformer.h.95.ln_mlp.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.95.mlp.dense_4h_to_h.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.95.mlp.dense_h_to_4h.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.95.self_attention.dense.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.95.self_attention.query_key_value.weight": "model-00096-of-00101.safetensors",
+    "transformer.h.96.ln_attn.bias": "model-00098-of-00101.safetensors",
+    "transformer.h.96.ln_attn.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.96.ln_mlp.bias": "model-00098-of-00101.safetensors",
+    "transformer.h.96.ln_mlp.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.96.mlp.dense_4h_to_h.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.96.mlp.dense_h_to_4h.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.96.self_attention.dense.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.96.self_attention.query_key_value.weight": "model-00097-of-00101.safetensors",
+    "transformer.h.97.ln_attn.bias": "model-00099-of-00101.safetensors",
+    "transformer.h.97.ln_attn.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.97.ln_mlp.bias": "model-00099-of-00101.safetensors",
+    "transformer.h.97.ln_mlp.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.97.mlp.dense_4h_to_h.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.97.mlp.dense_h_to_4h.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.97.self_attention.dense.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.97.self_attention.query_key_value.weight": "model-00098-of-00101.safetensors",
+    "transformer.h.98.ln_attn.bias": "model-00100-of-00101.safetensors",
+    "transformer.h.98.ln_attn.weight": "model-00100-of-00101.safetensors",
+    "transformer.h.98.ln_mlp.bias": "model-00100-of-00101.safetensors",
+    "transformer.h.98.ln_mlp.weight": "model-00100-of-00101.safetensors",
+    "transformer.h.98.mlp.dense_4h_to_h.weight": "model-00100-of-00101.safetensors",
+    "transformer.h.98.mlp.dense_h_to_4h.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.98.self_attention.dense.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.98.self_attention.query_key_value.weight": "model-00099-of-00101.safetensors",
+    "transformer.h.99.ln_attn.bias": "model-00101-of-00101.safetensors",
+    "transformer.h.99.ln_attn.weight": "model-00101-of-00101.safetensors",
+    "transformer.h.99.ln_mlp.bias": "model-00101-of-00101.safetensors",
+    "transformer.h.99.ln_mlp.weight": "model-00101-of-00101.safetensors",
+    "transformer.h.99.mlp.dense_4h_to_h.weight": "model-00101-of-00101.safetensors",
+    "transformer.h.99.mlp.dense_h_to_4h.weight": "model-00100-of-00101.safetensors",
+    "transformer.h.99.self_attention.dense.weight": "model-00100-of-00101.safetensors",
+    "transformer.h.99.self_attention.query_key_value.weight": "model-00100-of-00101.safetensors",
+    "transformer.ln_f.bias": "model-00101-of-00101.safetensors",
+    "transformer.ln_f.weight": "model-00101-of-00101.safetensors",
+    "transformer.word_embeddings.weight": "model-00001-of-00101.safetensors"
+  }
+}

config.json

@@ -0,0 +1,29 @@
+{
+  "_name_or_path": "tiiuae/falcon-180B",
+  "alibi": false,
+  "architectures": [
+    "FalconForCausalLM"
+  ],
+  "attention_dropout": 0.0,
+  "bias": false,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "hidden_dropout": 0.0,
+  "hidden_size": 14848,
+  "initializer_range": 0.02,
+  "layer_norm_epsilon": 1e-05,
+  "max_position_embeddings": 2048,
+  "model_type": "falcon",
+  "multi_query": true,
+  "new_decoder_architecture": true,
+  "num_attention_heads": 232,
+  "num_hidden_layers": 100,
+  "num_kv_heads": 8,
+  "parallel_attn": true,
+  "rope_scaling": null,
+  "rope_theta": 10000.0,
+  "torch_dtype": "float16",
+  "transformers_version": "4.36.2",
+  "use_cache": true,
+  "vocab_size": 65024
+}

generation_config.json

@@ -0,0 +1,6 @@
+{
+  "_from_model_config": true,
+  "bos_token_id": 11,
+  "eos_token_id": 11,
+  "transformers_version": "4.36.2"
+}

model-00001-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:19a739ec867d1bafac30d92029c17cc0cb79c28e28f24292374a5505754079b9
+size 4606919184

model-00002-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3470686d430581b905ac9e4594d919edd424b0256b3f181896c49f19d139919e
+size 4439790472

model-00003-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4e81f718dac89c93d958e78cab736a927669ce2eeb9444899ed98d2fd0c81ed3
+size 4439790472

model-00004-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:69cc020798153a1ccdc268bf3f8c4f0aac51f13e6137bfb7085e7ed682832d0d
+size 4439790472

model-00005-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:60ec8fa37abea842e6a3b062159c350fc88dadad9d71b6c75672140ba134e143
+size 4439790472

model-00006-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0028b5051ac988fed0a363f42e6336c9bf75b4f39890ca7f551bb04af3c2c86
+size 4439790472

model-00007-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fac029fb08ac9b7b4fede4f71ac012ffd5b5886d63452b445541590e495bc1af
+size 4439790472

model-00008-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:85dbc61d2b48858c861523ab07397a05a04d2d1fe597c1841fec5b6cd4522050
+size 4439790472

model-00009-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3fce6d8428f5f426fe9f68827c0aa92f92d7ea182206b5489e6004927d6f2775
+size 4439790472

model-00010-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7f4b011189d3776271aec0457b9881c54ba5b75984563598ec42fe09cabff3d6
+size 4439790472

model-00011-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f0727b179e63f2a339daa5365c520ac6d4215c26c9dff35fcf9be93c325f1b0c
+size 4439790512

model-00012-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0210cb4870f0f7de263092024ffc40c6b0d9538208adcdf4ae79c90d56d3be65
+size 4439790480

model-00013-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1cbd274bee0ad4afe32e0e3dbd2df6238cb459e2fddc46b0c791dbd7ab87ec92
+size 4439790480

model-00014-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:584a87abe5f3f309365f982b26520ff8bd4d0ea6cf56cb662a151995649792f8
+size 4439790480

model-00015-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:86dbc54ec71e257c08c1f010a03fde2f91a1b53a424979b0599f7336618cf66e
+size 4439790480

model-00016-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f620ba160734dd9d4603238477a5729dd73507e30d47de2347000a4b17a3ba27
+size 4439790480

model-00017-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1fc872e3976f46320dc20a19816d5cc6a531107e9525689765be32aab3fa29bb
+size 4439790480

model-00018-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2e983aecb65ea3e951bfaabf487a7cbb89ebe3e176801801c689a7830bb9f541
+size 4439790480

model-00019-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0eb6162225b27d8a763f854ee68dc1da92c2cba95169c1c0f7e3c97841553daa
+size 4439790480

model-00020-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9b70b2d99fbccd583ce9bbe0a7371aa6e687d185c99dc174c164d48e5ab572fd
+size 4439790480

model-00021-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:db750302e90c6e94c6f0d3171117ed8135aa724d78f9eb2847f90b55ec51d7f6
+size 4439790480

model-00022-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:489d309e42d85d2a2cb1109874a820ac536f960dcd47afb023694d76834cff6e
+size 4439790480

model-00023-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d88dee0ae69c2d59d4ac7f06e98659a9cd834202401cafdf9edc1fab2c8a9b6e
+size 4439790480

model-00024-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d363f967f99d719caf2728ff0861339227041d2358018f6edbdd559b06ff4476
+size 4439790480

model-00025-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:efae315915f2fa16ec29cd9cefd77d73dee77c607d29a745df584e56ac63c931
+size 4439790480

model-00026-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:58c15abe6c87a63e1d23faec0391094dc562a6e1d49202427f1758ce51d6c3aa
+size 4439790480

model-00027-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:15fbdb9a5133131014da10896441e9a0dae0636db54701031ac744909080a341
+size 4439790480

model-00028-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fdf21a9edccba705c4a814de5916b357d461b5f620086c97fd41e7ccf2d3895b
+size 4439790480

model-00029-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b26bc4e88a5e006a39b9cacc8027ff19e180a6b607d3731fb9de143bbf32b42d
+size 4439790480

model-00030-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9de57c7924773ea977d6ff12bd50bfca11c331cbb64f89a83a08945bf1aec72b
+size 4439790480

model-00031-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1570f4d759fc3c77c5328c435f6ecf7b4b85f53661d08c8170c5bb8b7b0b18db
+size 4439790480

model-00032-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:cb9832792d0e572027889615d84ae83a0b3c5912e8f0a95a18b9863ebc813526
+size 4439790480

model-00033-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b412dddcc7ca1036fe4768765a02ba66930047792b3443d12293b76c0a1bbfd4
+size 4439790480

model-00034-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:853b31249a2646f7b3a986d32e977d7cc013926ee489d23fff2dafded2b3d808
+size 4439790480

model-00035-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3be3c2bba5d1510337593fcf0fb9da3dfdb60ee44d3984ecc283aadf50c8e7be
+size 4439790480

model-00036-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2691c1c834f228504ded305a2ed69697900d699c5f136034a8cacda2846af240
+size 4439790480

model-00037-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2f989379f080ed2734b75300143233a8625489406e204baaffee04dfb6e494e2
+size 4439790480

model-00038-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4cf44956b4d2d635a8855824e0737381ece9618a5fd205dacb9ce9e6dcad5c8e
+size 4439790480

model-00039-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fcb5b017740cd7f74b78e1bc4ab9cdb2336d06fb304f4972bda3f26783d3f95d
+size 4439790480

model-00040-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:68dd17057dc0f258e76d78709e2aeed1b683e14b1344c12af5028fd247bff55d
+size 4439790480

model-00041-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d2f3cbaf486e3d3631b3075d0f72edc67afd7f92556212f50e26f54384a5b6b9
+size 4439790480

model-00042-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8e3a9e0f2b8a6a8929f293bf1bac53ffda150d19f110b949ede374b94334e778
+size 4439790480

model-00043-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c9e97fc11366cbcc576368cacb5ad9b285f846ee5cf4ceb73cfddbbc90955a96
+size 4439790480

model-00044-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1b3677cb368caae194a50a4e26706a3de369deb6b8fb1235eae4a6f817a51272
+size 4439790480

model-00045-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:3c81874e6dc754a169fb2f8cac64caca7d814b22029230ce3aaa8eaba686bd3f
+size 4439790480

model-00046-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:506f5171f0d40757142494d834754b439a48ee5724d713762c6b88cc44fcc98b
+size 4439790480

model-00047-of-00101.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:429d2bbafb42176809fa412178a8501598edd7dfacb85851358d98ec31a8feab
+size 4439790480